Environ. Sci. Technol. 2010, 44, 901–907
Deposition of Mercury Species in the Ny-Ålesund Area (79°N) and Their Transfer during Snowmelt ´ L I E N D O M M E R G U E , * ,† AURE C A T H E R I N E L A R O S E , † , ‡ , § X A V I E R F A ¨I N , | O L I V I E R C L A R I S S E , ⊥,# DELPHINE FOUCHER,⊥ HOLGER HINTELMANN,⊥ D O M I N I Q U E S C H N E I D E R , ‡,§ A N D CHRISTOPHE P. FERRARI† Laboratoire de Glaciologie et Ge´ophysique de l’Environnement, Universite´ Joseph Fourier - Grenoble 1/ CNRS, 54 rue Molie`re BP56, 38402 Saint Martin d’He`res France Laboratoire Adaptation et Pathoge´nie des Micro-organismes, Universite´ Joseph Fourier - Grenoble 1, BP 170, 38042 Grenoble Cedex - France CNRS UMR 5163 Division of Atmospheric Sciences, Desert Research Institute, 2215 Raggio Parkway, Reno, NV, USA Department of Chemistry Trent University 1600 West Bank Drive Peterborough ON K9J 7B8 - Canada Now at De´partement de chimie et biochimie Universite´ de Moncton Moncton, N.-B. E1A 3E9 - Canada
Received August 25, 2009. Revised manuscript received November 30, 2009. Accepted December 2, 2009.
Arctic snowpacks are often considered as temporary reservoirs for atmospheric mercury (Hg) deposited during springtime deposition events (AMDEs). The fate of deposited species is of utmost importance because melt leads to the transfer of contaminants to snowmelt-fed ecosystems. Here, we examined the deposition, fate, and transfer of mercury species (total Hg (THg) and methylmercury (MeHg)) in an arctic environment from the beginning of mass deposition of Hg during AMDEs to the full melt of the snow. Following these events, important amounts of THg were deposited onto the snow surface with concentrations reaching 373 ng · L-1 and estimated deposition fluxes of 200-2160 ng · m-2. Most of the deposited Hg was reemitted to the atmosphere via photochemical reactions. However, a fraction remained stored in the snow and we estimated that the spring melt contributed to an input of 1.5-3.6 kg · year-1 of THg to the fjord (i.e., 8-21% of the fjord’s THg content). A monitoring of MeHg in snow using a new technique (DGT sensors) is also presented.
Introduction Mercury (Hg) is a toxic metal that can be transported far from its emission sources and can contaminate remote areas. Anthropogenic sources of mercury (combustion, chemical * Corresponding author phone: +33 476 82 4211; fax +33 4 76 82 42 01; e-mail:
[email protected]. † Laboratoire de Glaciologie et Ge´ophysique de l’Environnement, Universite´ Joseph Fourier - Grenoble 1/ CNRS. ‡ Laboratoire Adaptation et Pathoge´nie des Micro-organismes, Universite´ Joseph Fourier - Grenoble 1. § CNRS UMR 5163. | Division of Atmospheric Sciences, Desert Research Institute. ⊥ Department of Chemistry - Trent University. # Now at De´partement de chimie et biochimie - Universite´ de Moncton. 10.1021/es902579m
2010 American Chemical Society
Published on Web 12/18/2009
industry, and mining) have significantly decreased since the 1970’s in the western countries of the northern hemisphere, but these declines are today certainly offset by increases in Asia (1). Whether the variations of sources have affected background concentrations of gaseous elemental mercury (Hg°), the predominant form of Hg in the atmosphere, is under discussion (2–4). Nowadays Hg° averages 1.5-1.7 ng · m-3 in the northern hemisphere. Despite the effort made to reduce anthropogenic emissions, levels of oxidized species in environmental matrixes remain elevated. Some studies even reveal an increase in Hg levels that can be seen in many species all over the planet including in some arctic animals (5). Of particular concern is the exposure to methylmercury (MeHg), a powerful neurotoxin that biomagnifies in the food chain. The Hg issue in the Arctic generates interest in scientific communities (e.g. ref 6) due in part to wildlife and human health implications. The reasons for elevated Hg levels in the Arctic have yet to be elucidated. Historically, the discovery made in Alert (Canada) in 1995 (7) that revealed that Hg° is oxidized and deposited onto polar environmental surfaces more rapidly than anywhere else due to a phenomenon called atmospheric mercury depletion events (AMDEs) sparked considerable interest. Significant increases in oxidized levels of Hg have been measured in the snow following AMDEs and a fraction of Hg has been reported as bioavailable (8). Yet the link to the methylation of Hg in this system is still unknown. The AMDEs were thought to be a dominant source of contamination to arctic ecosystems. Yet, an increasing number of studies seem to suggest that their contribution is not as important as previously reported. Among these, several studies have shown since 2002 that a part of deposited mercury can be reemitted back to the atmosphere by photomediated reduction processes (e.g. ref 9) thus adding uncertainty to the net deposition. A recent report even argues that the atmospheric input might not be as high as expected (10). Besides the ambiguous role of AMDEs in the contamination of the Arctic, the mechanisms that produce MeHg in these cold environments are to date unresolved although many pathways have been suggested (11). To fill some of the gaps highlighted in the literature and the still open questions, longer data sets and new analytical tools are required. The objective of the following study is to better constrain the fate of Hg once in the snowpack and its transfer to other reservoirs. For this purpose, we monitored mercury changes in the snow reservoir at a high temporal resolution, together with measurements in surrounding glaciers and in many meltwater streams. Along with THg, MeHg concentrations are reported. The study took place at the onset of AMDEs to the snowmelt of a seasonal snowpack in an arctic coastal location near Ny-Ålesund, Svalbard (Norway).
Materials and Methods Field Campaign and Study Site. The spring research campaign was held between April 16th, 2007, and June 20th, 2007, at Ny-Ålesund in the Spitsbergen Island of Svalbard, Norway (78°56′N, 11°52′E). The field site, a 50 m2 perimeter with restricted access (to reduce contamination from human sources), is located along the south coast of the Kongsfjorden, which is oriented SE-NW and open to the sea on the west side (Figure S1 of the Supporting Information). The Kongsfjorden was free of sea ice throughout the campaign. Atmospheric Hg° Measurements. Atmospheric Hg° was measured 2 m above snow surface using a Tekran model VOL. 44, NO. 3, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
901
FIGURE 1. (a) Atmospheric Hg° concentrations at sea level (black line) are shown on the left axis. Atmospheric Hg° data from the 16th to the 26th of April (cross) were obtained at Zeppelin air monitoring station (474 m above sea level) and are provided courtesy of NILU. (b) Hg° fluxes recorded from the snowpack (gray line) are shown on the right axis. Gray rectangles represent AMDEs. For clarity, AMDEs occurring between April 21st and April 27th are not shown. 2537A vapor-phase Hg analyzer. Because of logistical problems, measurements were available at our field site from April 26th. Data from the 16th to the 26th (courtesy of NILU) were at Zeppelin air monitoring station (474 m above sea level). An international study has already compared both measurement sites (12). Snow-to-air fluxes of Hg° were also continuously measured from April 26 in 10 min intervals using a dynamic flux chamber (13). Detailed characteristics and testing are available in SI. Snow and Seawater Sampling. Snow was sampled daily using clean sampling techniques and was collected in acidwashed 250 mL borosilicate glass bottles. Snow samples were also collected from pits dug in the surrounding glaciers (Figure S1 of the Supporting Information), Kongsvegen (N78°45.29, E13°20.20, 670 m) Austre Love´nbreen (N78°51.79, E12°08.03, 472 m) and Holtedahlfonna (N79°08.17 E13°16.12, 1173 m). All snow samples were maintained frozen in the dark until analysis. Surface seawater was sampled at a distance of 50 m from the shore (200 m from the snow field site) every 5 days. Meltwater Sampling. Meltwater was sampled using two different methods. The first consisted of a Teflon-coated v-shaped sampler that was placed at about 10 cm under the surface of the snowpack. This surface meltwater runoff was collected in clean 250 mL borosilicate glass bottles. Snowpack meltwater was collected in sterile sampling bags at the snow-soil interface as small rivers formed. Both collection methods were tested by multiple blanks measurements. THg concentration was measured constantly below detection limit in MQ water sampled on our Teflon-coated v-shaped sampler as well as in sterile sampling bags. Analytical Techniques. THg Determination. THg was measured in the field with a Tekran model 2600 using USEPA method 1631 revision E. All samples were analyzed in triplicate. THg concentrations are presented as mean (1 standard deviation. The limit of quantification calculated as 10 times the standard deviation of a set of 10 blanks was 0.3 902
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 3, 2010
ng · L-1. Detection limit (DL) is about 0.1 ng · L-1. Analytical information is available in SI. Methylmercury Determination Using DGT (Diffusive Gradient in Thin Films). Developed initially as a monitoring tool for MeHg in the water column (14, 15) or in sediment pore water, DGT technique was used for the first time in the snow and melting snow. Three DGT units were deployed every five days in the snowpack 5 cm below the surface. DGTs were also deployed in triplicate for 5 days at a depth of 1 m in the nearby fjord. DGT principle is based on the diffusion of the dissolved species through a gel and their subsequent accumulation on an ion-exchange resin. After extraction and elution in the lab, MeHg was ethylated and analyzed by GC-ICP-MS. DGT blanks, collected on a regular basis in the field, assessed the absence of contamination of this new monitoring tool. In water, interpretation of the DGT results in terms of a labile MeHg concentration is straightforward. In snow and melting snow, the calculated DGT MeHg concentrations are probably not absolute values, but reflect MeHg mobility, that is high DGT MeHg concentrations represent high MeHg mobility in snow or melting snow. Details on the technique the limitations and a comparison with a conventional MeHg determination are available in the Supporting Information.
Results Atmospheric Hg° Concentrations and Flux Measurements over Time. AMDEs, defined when hourly averages of Hg° concentrations fall below 1 ng · m-3, were complex and of short duration (part a of Figure 1). Two long AMDEs occurred: from April 18 to April 20 and from May 26 to May 28. Additional short and shallow AMDEs were recorded on April 21, April 22, April 24, April 27, May 1, May 4, and May 8. During this extended period (April 18 to May 8), Hg° concentrations oscillated between 0.6-1.0 ng · m-3 and background concentrations (∼1.5 ng · m-3). The beginning of the first AMDE (April 18-20) corresponded to a snowfall
FIGURE 2. THg concentration (black diamonds, left axis) in surface snow samples as a function of the day of the year. Samples were analyzed in triplicate. The error bars show the standard deviation of the 3 triplicates. Errors bars are generally less than 0.2 ng · L-1 and are thus not visible with the present scale. pH (squares, right axis) of melted snow samples is shown as well. A few triplicate measurements of THg in the nearby surface fjord water are also presented in the insert panel. event as well as an episode of strong wind coming from the Arctic Ocean on the west side of Spitsbergen. On the basis of the HYSPLIT results, air masses originated from Northeast Greenland, where high BrO concentrations in the atmospheric column were detected. The sea ice map shows that the air mass had passed over sea ice and open ocean areas (part a of Figure S2 of the Supporting Information). The AMDE was at its maximum amplitude (i.e., lowest Hg° values) on the 19th, when the wind speed dropped (1-2 m.s-1). Finally, Hg° concentrations returned to background levels (above 1 ng · m-3) on the 20th when wind from inland Spitsbergen swept away the depleted air and local reactants to replace it with air containing background Hg° concentrations. The second long AMDE (May 26 to May 28) occurred simultaneously with a temperature drop and during a snow fall. In fact, it started with the contribution of a cooler air masses coming down from the Kongsvegen glacier and a high wind speed (7-8 m.s-1). During this period, northeastern and eastern regions from Spitsbergen presented higher BrO concentrations. Backward trajectories show a contribution of air masses coming from the North Pole (part b of Figure S2 of the Supporting Information). Hg° Trends and Flux Measurements. Atmospheric Hg° concentrations were influenced by emissions from the snowpack. In the absence of AMDEs, a diel pattern of Hg° with higher variability and higher values was observed during the day, with values reaching 2.5 ng · m-3 periodically. Hourly averaged Hg° concentration even stayed above 3 ng · m-3 for 6 h (peaking at more than 4 ng · m-3) on May 29 after the last AMDE (part a of Figure 1). Flux measurements confirmed the importance of Hg° emission from the snowpack (part b of Figure 1). They revealed a clear diel pattern with daily hourly maximum averaging more than 20 ng · m-2 · h-1 (average of the whole campaign). Large emission events were observed following both AMDEs. Following the second AMDE, fluxes reached 1000 ng · m-2 · h-1 and atmospheric Hg° was influenced by snow reemission up to a height of 7 m (K. Aspmo, personal communication). Forced ventilation of
interstitial snow air, however, may have led to an overestimation of Hg° fluxes during this event. Excluding the high reemission event mean Hg° fluxes was around 12.5 ng · m-2 · h-1 for the whole campaign. Speciation of Hg in the Surface Snow. THg concentrations ranged from below detection limit to 373.1 ng · L-1 (Figure 2). THg increased as Hg° concentrations decreased, indicating Hg° oxidation and deposition. From the beginning to April 24, THg averaged 191 ng · L-1, due to deposition of oxidized species of mercury (Hg(II)) through AMDEs. Hg deposition was enhanced by the occurrence of snow precipitation. Between the evening of the 24th and the morning of the 25th, surface THg fell from 356 to 70 ng · L-1 due to a snow fall. Because almost 20 cm of snow covered the previous layers and because an AMDE had just terminated, the THg value of 70 ng · L-1 reflected a measurement of wet deposition of Hg(II). In the absence of Hg deposition events, the THg content in surface snow slowly decreased to values below 2 ng · L-1. The last AMDE at the end of May increased THg levels from 2.1 to 237.7 ng · L-1. Photoreduction of deposited mercury and evasion of Hg° occurred immediately after (part b of Figure 1). After 3 days, surface THg concentration had returned to “normal values”, that is that 99% of deposited Hg has been transformed, transferred, or reemitted. Methylmercury species (MeHg), as integrated over 5 day periods by DGT probes in surface snow, showed increasing levels throughout the campaign (Figure 4 and Table S1 of the in Supporting Information). A first period showed low levels averaging around 0.049 ( 0.022 ng · L-1. Then, after May 10, higher values and higher variability were observed (0.117 ( 0.050 ng · L-1). A maximum of 0.259 ( 0.015 ng · L-1 was recorded on June 3 to 8. Speciation of Hg in the Seawater. THg in the surface seawater close to the snow sampling area was quite elevated around 1.5 ng · L-1 (Figure 2). A net increase occurred on May 29 with a measured value of 2.9 ng L-1. DGT probes recorded MeHg values from 0.007 to 0.158 ng · L-1 (Figure 4 and Table S1 of the Supporting Information). The highest VOL. 44, NO. 3, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
903
FIGURE 3. THg concentrations in surface meltwater (dark-gray bars) and snowpack meltwater (light-gray bars) as defined in the method section. The standard deviation of the triplicate analysis of each sample is represented with error bars. Data of pH and THg of surface snow already shown in Figure 3 are also plotted. The first snowmelt sample was collected on May 24. From May 26 to May 31, a cooler period interrupted meltwater collection.
FIGURE 4. MeHg concentrations in (a) surface snow and (b) seawater over 5 days as measured by the DGT technique (Materials and Methods and Supporting Information). The central dot represents the mean value of 3 DGT deployed at the same time and at the same place. The rectangle height shows the standard deviation of the 3 DGT, and the rectangle length the duration of the MeHg collection by the DGT. The DGT do not show a point value of MeHg but rather an integration of 5 days of MeHg presence in the snow. Precipitation events are represented by bold black spikes (no scale). The snow temperature (15 cm below the surface) is shown on the right vertical axis. values were recorded on samples of May 19-24 and May 29 to June 3 (respectively 0.158 and 0.131 ng · L-1). THg on Surrounding Glaciers. The three pits were dug at different locations (Figure S1 of the Supporting Information) and different periods. As shown in Table 1, the 904
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 3, 2010
Kongsvegen pit (April 20) had elevated surface THg concentrations (30.3 ng · L-1); however, below this depth THg was comprised between 0.5 and 7.0 ng · L-1. High values were found at the bottom of the pit on the icy layer, which corresponded to the refrozen layer of the summer melt. The
TABLE 1. THg Concentration Measured in Pit Dug on Three Different Glaciers (Figure S1 of the the Snow Accumulation as of Last Summera glacier (date) Depth (cm) surface 25 ( 10 50 ( 10 75 ( 10 100 ( 10 125 ( 10 150 ( 10 175 ( 10 200 ( 10 225 ( 10 a
Kongsvegen (April 20) 31.0 ( 0.3 5.5 ( 0.1 2.4 ( 0.1 1.6 ( 0.1 0.7 ( 0.1 0.5 ( 0.1 3.7 ( 0.1 2.9 ( 0.1 3.5 ( 0.1 7.0 ( 0.1
Supporting Information
) Representing
Austre Love´nbreen (May 18) THg (ng · L-1)
Holtedahlfonna (May 15)
1.2 ( 0.1 57.4 ( 0.3 15.2 ( 0.1 0.4 ( 0.2 0.6 ( 0.4 2.4 ( 0.1 0.2 ( 0.1
